branchrust_reboot/main/ice_sculptures2/sdf_sculptingcancel
8 Commits over 0 Days - ∞cph!
reduced triangles->vertices/indices processing by roughly half by making Triangles/EdgeKeys readonly struct and only calculating the hash once on creation, done in parallel when they're created and reduces the overhead in the single-thread processing stage
improved triangles->indices generation performance a bit, mainly trying to make it somewhat readable
sped up marching cube triangle generation by 10x by using local list batches with IJobParallelForBatch to reduce NativeList<>.ParallelWriter locking adds
(~0.35ms -> ~0.05ms)
Indirectly iterate modifiers with separate list of modifier types that servers as history
trying to make sense of shit code I wrote a year ago
got test scene working properly
merge from ice_sculptures/sdf_sculpting
